Theoretically I think this could be possible. Twins who are genetically identical can be impossible to tell apart. This would suggest that it should be quite possible to tell what a person looks like from their genetic makeup.
The problem is turning theory into practice. We have sequenced the entire human genome but we are only just beginning to get an idea of how genes, and the many DNA sequences that modify gene expression, produce proteins – and then how these proteins are modified to produce things as complicated as facial features. There are likely to be many hundred genes involved in making a face, and there are likely to thousands of modifiers that affect these genes. Gene expression is likely to be subject to many interactions and feedback networks all of which can change what happens to a developing face in many many ways. Once proteins are made these too can be modified or affected by other proteins, cellular processes and environmental factors. It is going to take us many years of work to tease apart all of these theoretical pathways and interactions and some radical improvements in computer modeling before we can say with any degree of confidence that a certain DNA sequence is likely to result in a big nose.
The other thing to take into consideration is how age will affect facial features. Here environmental features are likely to play a big part in how a person looks. Did you break your nose? Do you smoke? These sorts of things can change how you look considerably.
So I think this is a long way off and even then it’s likely to be limited to probabilities rather than an absolute fit. I do however think it’s just a matter of time until pretty good likeness are produced….how much time? Who knows? Maybe we’ll need to wait until CSI Tranquility Base is on TV!